Experimental infection of Eurasian collared-dove (Streptopelia decaocto) with West Nile virus



The Eurasian collared-dove (Streptopelia decaocto) has recently experienced a population explosion in North America. It is frequently infected with West Nile virus (WNV). To test the hypothesis that the Eurasian collared-dove is competent to transmit WNV, we experimentally infected two cohorts of doves with two different strains of WNV, CO08, and NY99, respectively. Both virus strains induced a low-level viremia, capable of infecting a small fraction of vector mosquitoes. We suggest that the Eurasian collared-dove plays a relatively insignificant role as an amplifying host for WNV, but it may be important where it is locally abundant.


The Eurasian collared-dove (Streptopelia decaocto) has recently experienced a population explosion that has caused it to disperse from its original range of India, Sri Lanka, and Myanmar (Smith 1987, Gorski 1993, Romagosa and McEneaney 1999). In the last century it dispersed westward, colonizing virtually all of Europe north to the Arctic Circle in Norway and south to North Africa. It also moved east into China, as well as Japan where it was presumably introduced (Crooks and Soule 1999, Rocha-Camerero and Hidalgo de Trucios 2002, Eraud et al. 2007). In North America, these doves were first observed in the early 1980s in the Bahamas, where they likely escaped from captivity (Smith 1987). Subsequently, they appeared in Florida. The birds dispersed north and west from Florida by establishing small population foci in advance of colonized regions. This type of colonization is termed “jump dispersal,” where the areas in between are gradually filled in (Fujisaki et al. 2010). This pattern is similar to what has been observed on the European continent (Rocha-Camerero and Hidalgo de Trucios 2002). Its success in colonizing North America is largely due to its propensity for thriving around human settlements (Hudson 1972).

West Nile virus (WNV) has followed a similar range expansion within the last century. Its site of origination is unknown, but early detections occurred in Uganda in 1937, in the Middle East in the early 1950s, and shortly thereafter in Pakistan (Smithburn et al. 1940, Nir et al. 1968, Hayes and Burney 1981). By the mid-1990's, it had been detected throughout the Old World tropics, all of Africa, Europe, south Asia, and Australia (Kramer et al. 2007). It spread to New York City in 1999 and within a decade had invaded most countries of the New World (Komar and Clark 2006). It frequently infects doves, and several dove species are believed to serve as amplifying hosts (Kent et al. 2009). Although Lineage 2 WNV was isolated from this species in Italy (Savini et al. 2012), the relationship between WNV and the Eurasian collared-dove is uncertain.

In the western U.S.A., a region where the Eurasian collared-dove has become abundant (Beckett et al. 2007) and WNV is active annually (Lindsey et al. 2010), high WNV-specific antibody prevalence has been observed in these doves (CDC, unpublished data). Eurasian collared-dove may serve as an important amplifying host for WNV, facilitating the transmission and geographic spread of the virus wherever this expanding dove population invades. Alternatively, the doves may reduce WNV transmission rates if they prove to be incompetent hosts for amplifying the virus to its insect vectors, which typically are Culex mosquitoes. To examine these alternative hypotheses, we experimentally inoculated several Eurasian collared-doves with WNV and determined the magnitude and duration of resulting viremia, and the nature of the dove's humoral immune response.



Eurasian collared-doves (n=24) were collected in Wellington, CO, using 60 mm-mesh mist nets during March and April, 2010 and transported to the Division of Vector-Borne Diseases animal facility in Fort Collins. An initial 0.6ml blood sample was taken by jugular or brachial venipuncture from each individual bird to determine if WNV-reactive antibodies were present in their sera. A third of the Eurasian collared-doves captured in northern Colorado tested positive for prior flavivirus infection, with serum (diluted 1:10) neutralizing WNV by at least 90%. The remaining 16 antibody-negative doves were divided into two equal cohorts and infected with a Colorado strain of WNV from 2008 and a New York strain of WNV from 1999, respectively. Birds were cared for in accordance with an approved institutional animal use protocol. Cohorts of eight birds each were housed in two approved cages separated by metal shelving.

Experimental infection

Two cohorts of eight seronegative birds were needle-inoculated subcutaneously on the breast with 1,500 pfu/0.1 ml for WNV strains CO0813834 and NY994132, respectively. Both viruses had been passaged twice in Vero cells prior to use. Immediately prior to inoculation, each bird was exposed to a colony-derived, probing Aedes albopictus mosquito on the breast in order to simulate the natural conditions of inoculation by mosquito bite. All but three birds were euthanized by cervical dislocation after being anaesthetized with a mixture of ketamine hydrochloride (Ketaset) and Xylazine (Rompun) at 14 days post-inoculation (DPI). The remaining three birds were euthanized at 210 DPI.

Blood sampling regimen and procedure

Once infected, a blood sample (0.1 ml) was collected daily from each bird for 7 d for determination of viremia profiles. Blood was drawn by jugular or brachial venipuncture using disposable 27g ½-inch Sub-Q needles attached to 1 cc syringes. Whole blood was placed in microtainer® serum separator tubes, allowed to coagulate for 15–30 min at ambient temperature, and then centrifuged at 10,000 rpm for 3 min for separation of serum. The samples were then frozen at −80° C until thawed for viremia measurement. At 14 DPI, 0.6 ml was drawn for antibody detection. For the three doves held beyond 14 DPI, 0.6 ml blood was drawn at 28 DPI, and monthly thereafter.

Reservoir competence

Reservoir competence was calculated using the formula C = s * i * d, where s is susceptibility, the proportion of birds infected as a result of exposure; i is mean daily infectiousness, the proportion of exposed vectors that become infectious per day; and d is duration of infectiousness, the mean number of days that an avian host maintains a viremia (Komar et al. 1999). These data were produced using infectious viremia standard curves with a threshold level of 104.7 and 104.4 PFU/ml serum, for infection of Culex pipiens L. (y = 1349x–6235) and Cx. tarsalis Coq. (y = 2634x–1156), respectively, as a function of WNV viremic titer (Goddard et al. 2002, Turell et al. 2002, Kilpatrick et al. 2007).

Viremia measurement

Viremia was measured and calculated in terms of plaque-forming units (pfu)/ml using Vero plaque assay titrations based on serial ten-fold dilutions of serum (starting at 1:10) in BA-1 diluent (Hanks M-199 salts, 0.05 M Tris, pH 7.6, 1% bovine serum albumin, 0.35 g/liter of sodium bicarbonate, 100 units/ml of penicillin, 100 µg/ml of streptomycin, 1 µg/ml of Fungizone), in duplicate in six-well cell culture plates inoculated with 0.1 ml of diluted serum. After 1 h incubation at 37° C, cells are overlaid with 3 ml 0.5% agarose in DMEM media supplemented with antibiotics, incubated for three to four days at 37° C, 5% CO2, and overlaid again. The second overlay contains neutral red to permit plaque visualization. Plaques were counted one to two days after staining.

Antibody titer measurement

Neutralizing antibody titers from the three remaining doves were measured using a plaque-reduction neutralization test (Beaty et al. 1995). Serial two-fold dilutions of serum (starting with 1:5) were mixed with an equal volume of BA-1 containing approximately 100 pfu/0.1 ml WNV (strain NY994132). After 1 h incubation at 37° C, mixtures were inoculated onto Vero cell monolayers in duplicate wells of a six-well culture plate and treated the same as a plaque assay. Wells in which the numbers of pfu were reduced by 90% were considered positive and used to determine the PRNT90 titers.

Statistical analysis

To test for statistically significant differences among peak viremia titers, titers were log-transformed, and the means were compared using the Student's t-test.


Avian viremia

All doves developed viremia without signs of illness. Peak viremia for both virus strains occurred at 2 DPI with the maximum mean viremias reaching 105.0 and 104.6, respectively (Figure 1). The range of response for the individual subjects in each cohort was from 103.1 to 105.6 for the Colorado strain and from 102.5 to 105.5 for the New York strain. Mean viremias in both cohorts declined quickly after 2 DPI, becoming undetectable after 4 DPI and 5 DPI, respectively. Viremia profiles did not differ significantly between strains (Figure 2).

Figure 1.

Viremia profiles for Eurasian collared-doves infected with CO08 and NY99 strains of WNV.

Figure 2.

Comparison of viremia profiles (NY99 strain) of Eurasian collared-doves (EUCD), rock pigeon (ROPI, data from Komar et al. 2003) and mourning dove (MODO, data from Reisen et al. 2005).

Development of neutralizing antibodies

Three doves developed an antibody titer of ≥80 by 14 DPI and by 30 DPI all had antibody titers ≥160. Over the 210 day period, the titers varied very little and never dropped below 160 (Figure 3). Dove “EUCD 63” died on 102 DPI so no results are shown past 90 DPI.

Figure 3.

Antibody response of three Eurasian collared-doves (EUCD) from 14 dpi to 210 dpi as determined by PRNT.

Reservoir competence

Reservoir competence index values for collared-doves were calculated for two principal vectors of WNV, Cx. pipiens and Cx. tarsalis, and compared to other dove species (Table 1). Eurasian collared-doves were weakly competent for Colorado and New York strains of WNV.

Table 1. Comparison of reservoir competence index for 3 dove species:, Eurasian collared-dove (EUCD); rock pigeon (ROPI);, mourning dove (MODO). Index values are calculated with respect to two mosquito species, and two virus strains.
  1. *95% confidence interval.

Cx. pipiens    0.03 [0.00—0.14]*    0.00 [0.00—0.05]    0.00 [0.00—0.00]    0.03 [0.00—0.21]
Cx. tarsalis    0.14 [0.00—0.49]    0.01 [0.00—0.14]    0.00 [0.00—0.01]    0.34 [0.00—1.12]


Eurasian collared-doves have become increasingly common in the United States and in certain environments they could play a role in the amplification of arboviruses such as WNV (Beckett et al. 2007). Indigenous dove species in the U.S.A., such as the mourning dove (Zenaida macroura), have been shown to be weakly competent reservoir hosts for WNV in experimental infection studies (Komar et al. 2003, Reisen et al. 2005) and our results reveal similar viremia profiles for the Eurasian collared-dove, for both the original North American strain, NY99, and a more recent isolate, CO08. The viremia levels are sufficient to infect a small fraction of the principal Culex vectors that might feed on them, as indicated by the relatively low WNV reservoir competence index values.

The importance of the Eurasian collared-dove as an amplifier of WNV would depend on the local ecology of a transmission focus. The shear abundance of the species could enhance virus amplification where and when more suitable reservoir hosts (including passerine species such as the American crow [Corvus brachyrhynchos], blue jay [Cyanocitta cristata], American robin [Turdus migratorius], and house sparrow [Passer domesticus]) are unavailable. Furthermore, the low reservoir competence of doves may be offset by high vector contact, resulting in a low-level amplification of WNV. For example, in northern Colorado, the proportion of vector blood meals derived from the Eurasian collared-dove was 19% at the peak of the WNV transmission season in August (Kent et al. 2009). Avian surveys conducted as part of the study found that doves generally comprised a lesser proportion of the available bird population, indicating a strong host selection for these doves by Culex tarsalis, the local WNV vector in northern Colorado. Whether these vector-dove contacts contribute to virus amplification or virus extinction would depend on the immune status of the local dove population.

All three birds used for immune response monitoring developed virus-neutralizing antibodies by day 14 as expected with very little fluctuation in antibody titer over the 210-day period. This suggests that although viremia levels are low in this species, infected birds will have lifelong immunity to WNV. Long-lasting humoral antibody responses to WNV have also been observed in the rock pigeon (Columba livia) (Komar et al. 2003, Gibbs et al. 2005). Long-lasting immunity against future WNV infections would reduce the importance of these doves as amplifiers, because immune birds are incompetent as reservoirs. Even WNV-naïve doves may also effectively reduce local WNV transmission if the net numbers of infectious vertebrate-vector contacts are reduced by the presence of these doves. This could occur when and where the doves divert potentially infectious mosquito bites from highly competent vertebrate reservoir hosts, such as the passerine species mentioned earlier.

Field data on the immune status of local bird populations and vector-vertebrate reservoir contact rates for Eurasian collared-doves vs other avian species will be needed to determine the location-specific impact of this emerging dove population on local WNV amplification. The reservoir competence data presented in this paper make such a field study possible. The viremia profiles of Eurasian collared-dove chicks are also needed. Higher virus titers in the blood of nestlings could enhance amplification where vectors are feeding on defenseless chicks. Vector contact with collared-dove nestlings may enhance WNV amplification late in the WNV transmission season after the passerine breeding season. Unlike passerines, which brood from May to July in temperate latitudes, Eurasian collared-doves brood from February to October (Romagosa 2002), thereby potentially lengthening the transmission season where they are present. However, maternal transfer of humoral immunity to nestlings likely occurs and may reduce the importance of an amplification role for nestlings. More study on the role of nestlings in the transmission of WNV is needed.